A practical 13-week planning system with committed, conditional and learning work, leading indicators, decision thresholds and weekly reallocation rules.
Documented fact: Evaluation guidance supports defining questions, evidence and decision use in advance rather than judging work solely from post-hoc outcomes.
CDM inference: The three-lane 13-week system and its governance cadence are CDM editorial design, intended to make uncertainty visible rather than to forecast demand perfectly.
Why use a 13-week horizon?
Thirteen weeks is long enough to see creative, procurement, production and sales dependencies, but short enough to update as demand, price or capacity changes. It is an operating horizon, not a replacement for strategy, annual commitments or a cash forecast.
The plan should sit beneath the market thesis and above individual campaign boards. Strategy determines the problems and positions worth pursuing. The rolling plan shows what the organisation will commit, what remains conditional and what it must learn before the next allocation.
Do not confuse responsiveness with constant change. A volatile dashboard can tempt teams to rewrite priorities every Monday. The Magenta Book’s emphasis on pre-specified questions and proportionate evaluation is useful here: decide what evidence matters before the result arrives.
Use three planning lanes
| Lane | Purpose | Entry rule | Change rule |
|---|---|---|---|
| Committed delivery | Work the organisation has approved and must execute reliably | Clear owner, funded scope, dependencies and expected outcome | Change only at a gate or for material safety, legal, capacity or evidence failure |
| Conditional investment | Work ready to launch if a named condition is met | Trigger, expiry, minimum evidence and reserve budget are explicit | Activate, revise or release funds at the trigger date |
| Learning work | Tests that reduce a decision-critical uncertainty | Question, method, sample, guardrail and next decision are specified | Stop when the decision threshold is met or learning value falls |
A campaign may contain all three lanes. Core always-on service information can be committed; incremental media can be conditional on stock; message testing can sit in learning. Separating them exposes which spending is an obligation and which is an option.
Protect enough unallocated capacity to respond to evidence. This is not a universal percentage. It depends on contractual commitments, production lead time, category volatility and the cost of delay. Record the rationale instead of adopting a fashionable ratio.
Build the plan from decisions backwards
Begin with no more than three portfolio outcomes for the quarter. For each, name the customer or market mechanism, lagging outcome, leading indicators and guardrails. Add the decisions that must be made during the horizon: renew a partnership, expand a market, change a value message or release media reserve.
CDM asset: 13-Week Line Card
- Work item and lane
- Decision it supports
- Owner and accountable approver
- Cash cost and internal capacity
- Earliest start, lead time and point of no return
- Leading indicator, denominator and data owner
- Guardrail and stop condition
- Decision date and reallocation destination
Map dependencies visibly. A product feed cannot launch before identifiers are reconciled; a claim campaign cannot publish before evidence and legal review; a creator asset cannot enter paid media before usage rights are clear. A plan without dependencies is a wish list.
Add demand scenarios only where they change action. Use a base case, an upside condition and a downside condition with observable triggers. Avoid pretending each scenario has a precise probability unless the method supports it.
Finally, roll the horizon. At the end of every week, add a new week thirteen with provisional work and known decisions. This keeps future dependencies visible and prevents the team from consuming all capacity inside the current month.
Run a weekly evidence and exception review
The weekly review is not a presentation of every metric. Start with exceptions: which assumptions changed, which triggers fired, which dependencies slipped and which guardrails moved. Require the owner to propose a decision, not merely describe performance.
- Confirm changes in market, price, stock, capacity and regulation.
- Review leading indicators against their expected sequence.
- Inspect implementation quality before blaming the thesis.
- Decide conditional releases, stops or revisions.
- Record the decision, owner, evidence and next review date.
- Add the new week thirteen and resolve its dependencies.
Keep a decision log. If a team changes a campaign because qualified demand weakened, record the metric definition, comparison period and alternative explanations. Later learning should evaluate the decision using the information available at the time, not only the eventual result.
Set reallocation rules before pressure arrives
For every conditional line, specify where released money or capacity goes. Options include the next qualified experiment, a proven programme below saturation, customer research, product evidence or return to the business. “Spend it somewhere before quarter end” is not a reallocation rule.
Use thresholds as prompts for judgement, not automatic truths. A two-week signal may be sufficient for a reversible creative rotation and inadequate for terminating a long-cycle market programme. Match the window to the mechanism and decision cost.
Separate underdelivery from underperformance. If a platform spends half the planned amount, outcome variance may reflect delivery. If implementation reached the wrong segment, the test did not evaluate the intended mechanism. Finance, operations and marketing should agree these labels in advance.
Limits and risks
A rolling plan can become short-termism if every investment must prove itself within thirteen weeks. Brand memory, capability, distribution and product change may require longer horizons. Show those commitments and their nearer evidence milestones without inventing immediate revenue proof.
Frequent reallocation can create production waste, partner instability and noisy tests. Use gates, not constant optimisation. Legal, safety or misleading-content issues are exceptions and should trigger immediate correction.
CDM verdict
A 13-week plan should make uncertainty governable. Separate commitments from options and learning, define decision evidence before launch, review exceptions weekly and roll the horizon continuously. The plan succeeds when it improves the quality and speed of consequential decisions—not when every box stays unchanged.
Related guides
Frequently asked questions
Does a 13-week plan replace the annual marketing plan?This article is editorial guidance. Apply the principles in proportion to your market, evidence, and responsibilities.
